Re: Renaming GNAT? (was Re: Ada to C convertor)

Re: Renaming GNAT? (was Re: Ada to C convertor)

[Robert Dewar]

Fergus asked

"No?  What about 'unrestricted_access and such like?
Don't they constitute a distinct dialect?"

The Ada 95 LRM specifically allows an implementation to add impl defined
attributes (and pragmas).

But GNAT does NOT have any extensions other than such impl defined
features specifically allowed by the RM.

To find out if your program uses impl defined attributes, turn on the
features analysis switch -gnatx9, you will get a full cross reference
of any use of implementation dependent attributes or pragmas.

Re: Renaming GNAT? (was Re: Ada to C convertor)

[Robert Dewar]

Jon Anthony said

"I don't understand how this is true.  Since Gnat is under FSF
copyleft, and its sources are easily obtainable, and part of the
_intent_ of it was that universities and others would have a readily
available Ada compiler to do language experiments with, I believe that
it does indeed fit this perspective.  Now clearly that is _not_ what
will happen to the standard adhering, ACVC validated evolving versions
of it maintained by ACT, but that seems a different point."

Yet it is true. Surely people can play with extended versions of the
language in their own copies, and if they like distribute them, but
at least so far, the version maintained by ACT with the FSF copyright,
is exactly Ada 95, and NOT an extension of it. 

Re: Renaming GNAT? (was Re: Ada to C convertor)

[Jon S Anthony]

In article <dewar.824577635@schonberg> dewar@cs.nyu.edu (Robert Dewar) writes:

> Jon Anthony said
> 
> "I don't understand how this is true.  Since Gnat is under FSF
> copyleft, and its sources are easily obtainable, and part of the
> _intent_ of it was that universities and others would have a readily
> available Ada compiler to do language experiments with, I believe that
> it does indeed fit this perspective.  Now clearly that is _not_ what
> will happen to the standard adhering, ACVC validated evolving versions
> of it maintained by ACT, but that seems a different point."
> 
> Yet it is true. Surely people can play with extended versions of the
> language in their own copies, and if they like distribute them, but
> at least so far, the version maintained by ACT with the FSF copyright,
> is exactly Ada 95, and NOT an extension of it. 

Let me get this straight.  It _seems_ then that you are saying that
the term "GNAT" only applies to the FSF copylefted Ada95 compiler now
maintained by ACT.  And that the sort of experimentation I mention is
just fine, but the result could not be termed/called GNAT.  Is this
right?  Or is there something I am still not getting?

/Jon
-- 
Jon Anthony
Organon Motives, Inc.
1 Williston Road, Suite 4
Belmont, MA 02178

617.484.3383
jsa@organon.com

Re: Renaming GNAT? (was Re: Ada to C convertor)

[Tor Arntsen]

In article <dewar.824161921@schonberg>,
	dewar@cs.nyu.edu (Robert Dewar) writes:
>GNU Ada is potentially confusing, because GNU C is the name of a language
>(it is the extended C language compiled by gcc), and there is definitely
>no distinct GNU dialect of Ada.

I don't agree with this.  Most of us thinks of, and speaks of, GNU C as
a very good ANSI C compiler.  We do know that it has some extensions here
and there, but that doesn't make us think that 'GNU C' is the compiler
for the 'GNU C language'.  Absolutely not, in my opinion.

>It's OK to use GNU Ada informally, but you should be aware of this
>possible confusion. The GNU C phiolosophy is that it is fine to extend
>C with useful stuff. This is NOT a part of the gcc philosophy that
>GNAT copies.
>
>Anyway, as I noted earlier, GNAT is FAR too embedded at this stage for
>us to thing of changing it (and we like the name!)

I don't agree with this either, actually.  It may look like that for you, but 
you have lived within GNAT for a long time.  The rest of the world have no 
idea of what GNAT is, and how could they ever guess it's an Ada compiler?
Call it GNU Ada and they will know..

Regards,
-- 
Tor Arntsen (tor@spacetec.no)                    Standard disclaimers apply.

Re: Renaming GNAT? (was Re: Ada to C convertor)

[Richard A. O'Keefe]

tor@spacetec.no (Tor Arntsen) writes:

>I don't agree with this.  Most of us thinks of, and speaks of, GNU C as
>a very good ANSI C compiler.

I strongly disagree with this.  There is a _language_ called GNU C, which
has numerous extensions over standard C.  There is a _program_ called GCC,
which compiles that language (and several others).  This is the first time
since gcc was released that I've heard of anyone calling gcc gnu c.
(If you check the gcc man page carefully, you will find that the phrase
"GNU C compiler" most naturally admits the reading "compiler for the GNU
C language", given that the rest of the paragraph calls the program
"GNU CC".  Perhaps I should stop calling it gcc and call it GNU CC.)

>We do know that it has some extensions here
>and there, but that doesn't make us think that 'GNU C' is the compiler
>for the 'GNU C language'.

But there _is_ a "GNU C" language (it has nested functions, for example,
which C lacks), and gcc (or GNU CC) compiles it (amongst other languages).

I think most people would (and arguably _should_) regard the "name" of a
program as being the word they use to invoke it.  So I use C compilers
called "cc", "gcc", and "lcc", and as far as I'm concerned, these nouns
are their names.  If I call someone "Fred!" and he responds, then in that
context Fred is his name.  I do NOT invoke GNAT by typing "GNU Ada", so
"GNU Ada" does not _function_ as its name.  Most of the time, the name I
use for GNAT is "gnatmake".

>I don't agree with this either, actually.  It may look like that for you, but 
>you have lived within GNAT for a long time.  The rest of the world have no 
>idea of what GNAT is, and how could they ever guess it's an Ada compiler?
>Call it GNU Ada and they will know..

If it's called "GNU Ada", people will think it is a distinct dialect, just
as GNU C is not identical to C, GNU Pascal was not identical to ISO Pascal
Extended when last I looked at it, GNU C++ didn't accept the same language
as other C++ compilers (well, no two C++ compilers do, do they).

There are many people to consider:
1. the people who wrote GNAT
2. the people who maintain GNAT
3. the people currently using GNAT
4. the people who would use GNAT if they GNEW about it.
I find it very difficult to believe that the people in group 4 are being
or will be seriously confused by the name.  After all, where in the name
"UNIX" or "Linux" does it say it's an operating system?

-- 
Election time; but how to get Labor _out_ without letting Liberal _in_?
Richard A. O'Keefe; http://www.cs.rmit.edu.au/~ok; RMIT Comp.Sci.